[PATCH] [X86, AVX] use blends instead of insert128 with index 0

[Sanjay Patel]

Patch updated based on feedback from Andrea: for integer ops on AVX1, it's better to generate a wrong domain vblend instead of a wrong domain vinsertf128.


http://reviews.llvm.org/D8366

Files:
  lib/Target/X86/X86ISelLowering.cpp
  test/CodeGen/X86/2012-04-26-sdglue.ll
  test/CodeGen/X86/avx-cast.ll
  test/CodeGen/X86/vector-shuffle-256-v32.ll

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D8366.22282.patch
Type: text/x-patch
Size: 8621 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150319/07499b2d/attachment.bin>